Revision controlled Perl DataStructure

[mvr707]

I'm maintaining a directory as a perl hash. New employees get added and old get deleted as time passes by. What's the optimal way to maintain the hash such that I can recover the state of the hash at any given time?

Originally posted as a Categorized Question.

[rob_au]

Your best method of implementing this would be to write the data structure to a dated text file using a serialisation/dumper module such as Storable or Data::Dumper - An example piece of code might look something like this:

use Data::Dumper qw/Dumper/;
use IO::File;
use strict;

my %hash = {
    #   Your data structure lies within %hash
};

my $fn = "dump." . sprintf("%04d%02d%02d", (localtime)[5] + 1900, (loc
+altime)[4] + 1, (localtime)[3]);

my $fd = IO::File->new($fn, "w");
if (defined $fd) {
    print $fd Dumper(\%hash);
    $fd->close;
};

Note that this code is very rough and should not be used in this form - A better revision would incorporate file locking, tests for file existence and better time stamping.

 

perl -e 's&&[e-mail]&&&split/[@.]/&&s&.com.&_&&&print'

[expertseries]

Here's another, rob thanks for that example of sprintf.

print "Content-Type: text/plain;\n\n";
#------------------------------------------------------------------

# Current time as 'YYYYMMDD.HHMMSS' (GMT)
($sec,$min,$hour,$mday,$mon,$year,$wday,$yday,$isdst)=gmtime(time);
$revision = sprintf("%04d%02d%02d.%02d%02d%02d",
                    $year+1900,$mon+1,$mday,$hour,$min,$sec);

#------------------------------------------------------------------
print "\n\nRevision: $revision\n\n";

[mvr707]

But the idea of a general framework should be good audit tool on persistent datastructures.. Exactly like a revision control system like CVS allows me get a file as of any date/time...i am looking for similar feature for perl datastructure like hash which can be made persistent using Storable e.g.

Originally posted as a Categorized Answer.

[mvr707]

I am doing a similar stuff using Storable. But the strategy itself is not optimal... storing the FULL data for each snapshot. I was thinking of incremental storage and inheritance etc..

Originally posted as a Categorized Answer.